Retrospective Optimization of Daily Insulin Therapy Parameters: Control Subject to a Regenerative Disturbance Process**This material is based upon work supported in part by the National Institutes of Health under Grant No. NIH/NIDDK R01DK051562. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the sponsors. S. D. Patek and M. D. Breton disclose equity interest in TypeZero Technologies, Inc.

Abstract

This paper develops a novel approach to data-driven optimization of insulin pump treatment parameters in Type 1 Diabetes (T1D). In this approach, records of continuous glucose monitoring (CGM), insulin delivery, and meal records are used (i) to retrospectively estimate samples of the disturbance process that is responsible for daily variability in blood glucose and (ii) to optimize the parameters of functional insulin therapy (i.e. the patient's basal rate, correction factor, and carbohydrate ratio profiles) against the ensemble of estimated disturbance process samples. We illustrate the proposed methodology through retrospective application to data collected in a 30-day field study of patients with T1D, as well as through in silico pre-clinical trials using the FDA-accepted Virginia / Padova Type 1 Simulator.